Any out-of-the-normal phenomenon offers the opportunity for observation to learn some new things. Sometimes, new perspectives of old things emerge. Today, I was caught up in heavy traffic. Reason: Munima (the local hurricane)

First, I should have expected it. You know, we don't expect it when things happen. (other way round?)

One is always drawn to the ethics and unstated rules that the traffic system throws up. Almost, always they remain good on paper. When there is a traffic jam, any really objective methodology would suggest you route the packet through an alternate route. But you know, humans are humans. We would want to ensure that there is not even a small centimetre of road (a.k.a sandhu gap or cycle gap) through which one can sneek through. Only in this case, would one try to look for alternatives. In this respect, I find that the "dumb" routers exhibit greater intelligence than the "civilized and aware" human.

One contrast between the communication engineering problem and the traffic problem is that, in the latter, we have intelligent agents which act as signals. But there is one huge interference that guarantees that, in many cases, there is no solution - a million egos!

Dumb things are always sane. Humans aren't built to work that way. This, I feel is an important difference. If you notice successful problem solvers/solving groups, you will find that they have global sanity, but exhibit local insanity in a selective manner. As a small example, say, a human wants to walk from one end of a road to the other. A computer would follow the procedure that it was programmed to which is limited by the way in which the programmer perceives and understands the problem.

The human doesn't do it that way. He exhibits certain mannerisms, which we call eccentricities or individualities. One almost always, never misses the beautiful scenery that passes by. The existence of such individualities is due to the multi-dimensional optimization functions, that we probably employ. Exploring new spaces and ideas, I feel, is crucial in learning. So, is it just a question of bridging this gap? Ummm.. I don't know. May be, yes. But it is not an easy task considering the size and diversity of the representation vectors, the human brain seems to use.

---

Coming back to the uncivilized nature of our junta (I take myself to be part of it) - In the midst of all the commotion, I was thinking of this. The problems of trust, sharing and mutual respect replicate themselves at every place where there is an interaction between players. People have already modeled this in the form of Game Theory in which each player has his own cost function to maximize and there is a group, in this case, the whole ensemble of traffic, whose cost function is to be maxmized. Ofcourse, most of you should remember the Nash Equilibrium for zero sum games where the loss of one person is perfectly translated as a gain of another.

But, the problem so far with Game Theory(GT), has been its application. Very few people know the theory. Even fewer, know how to apply it. (I belong to neither of the above groups) Recently, the FCC in USA, has sanctioned the use of game theory in spectrum sharing. That it has already been applied to auctions, out of court settlement of disputes, market research is history now.  Ofcourse, resource allocation in CDMA networks is one thing I have seen often. Yes, I hear you. General resource allocation problems can be modelled as games. Network time, bandwidth, QoS and all those stuff. Nobody knows whether GT was "really" successful as to most of these problems, we don't have really good/mathematically closed solutions, as far as I've heard.

---

Coming back to the intelligent signals approach, I had mentioned... One would be forced to ask,"Uh.. Well, what do you mean? How can signals be intelligent?" How about "a chip that carries information" from one place to another, instead of an electromagnetic wave? How about intelligent molecule assemblages that travel from place to place conveying and taking information? This idea was motivated by the following problem.

---

The Railway Subway Channel and some observations:

In a railway subway channel, we have two inputs, two outputs. The global cost function would be the no. of channel uses per size of the subway or the average delay encountered. (more thought - capacity: the system definition or the signal definition?) Here are a few observations about this channel:

1) First, the subway was just hollow.People seemed to regulate themselves. But then, local intelligence need not (most often, does not) translate into global intelligence. This is why I sometimes I don't try to act too smart in a public place. Ofcourse, there are the  violators , who act as interference to others. Almost, everyone does. But these are extraordinary cases. They have little respect for others' spaces and sharing space. (more thought reqd: acceptable/unacceptable interference. conscious/sub-conscious/unconscious interferences. relation to the global interference functions)

2) The governing body, probably identified this and put a barrier in between, a sort of a median. Now signals moving in each direction would regulate themselves in their proper channel.

3) There is one disadvantage to (2). It doesn't take into account the distributions of the people in each direction.(I1->O1, I2->O2, I1->O2, I2->O1) But global intelligence does rise sometimes. During asymmetric traffic, people from the direction in which greater traffic comes, violate the normal rule of using only their side of the channel. They spill over to the unused portion of the opposite channel. (See the analogy to selective local insanity, above)

4) I hope you know how a railway subway (in India) is. Traffic flows: I1-O1, I1-O2, I2-O2, I2-O1. So one can also devise a better solution to avoid the collision between people travelling I2->O1 and I1->O2. But those are trivial problems, I guess.

(3) asked me a question. If signals are intelligent, they would regulate themselves? Yes. This is what is missing from the current communication problem. I don't see the concept of intelligent signals take root before advances in other basic sciences as nano-technology/vlsi systems. But I just hope there is scope in the not so far-off future.

One broader question: How different is this from the current Multiple Input Multiple Output systems, other than ofcourse the concept of intelligent signals? In MIMO systems, one antenna's signal received at the other is interference. In this case, it is not so. Transmission through all 4 channels is necessary and the sum constitutes the target maximization function. In normal MIMO systems, the maximization is that of rate of information transfer from the composite transmitter to the composite receiver. (more thought - see the analogy to the comparison between signal definition and system definition of  capacity)

---

To my friends: this has nothing to do with my current work. this is my independent, freelance thinking arising out of more time, I've got these days.
P.S: This is a really bad draft. Thanks for your comments/questions to refine this piece.